3
The Brotherhood of Coconuts
Unity, Conflict, and Narrowing Loyalties

Undugu wa nazi hukutani chunguni: [The] brotherhood of coconuts comes about in [the] cooking pot

This Swahili proverb wryly notes the difficult, even hopeless, circumstances under which coconuts unite.^[1] The unity of the Swahili community's major constituent parts came about through a conflict that shattered a long-enduring community organization and replaced it with one that is socially less inclusive and culturally narrower in scope. Following is a sketch of the events whereby the old organization ceased to operate and the newer one began to function.

Introduction

The decline of the Swahili community was particularly evident in members' beliefs about the quality of community life. My informants are unanimous in their belief that Twelve Tribes community life is not what it once was. Older people report that in their youth, neighbors and relatives were more dependable and the community as a whole more active. Younger people take the same basic view but refer, surprisingly, to the impropriety of their own behavior and the failure of their parents and respected community members to control them "better." I will return to the young people's view of their own conduct in chapter 4 when I consider the nature of culture's contents, but for now, the point is that informants of all ages agree that community life has declined.

Older informants sometimes mention the number, scale, and quality of group rituals—weddings, circumcisions, and funerals—that were held until fairly recently and the very substantial sums of money that were spent on

them. Men are quite ambivalent about this expenditure and sometimes say that the failure of the Swahili to keep up financially with their Hadhrami and Indian Old Town neighbors is due to the extravagance of the rituals held.

At the same time, there is a definite pride in their opulence. For women, this pride is particularly strong. There were still large weddings in the late 1970s, and large funerals were still held, although with lesser expenditures, throughout the 1980s. Outright regret from men at the diminished number and quality of rituals is only rarely heard, although it is heard from women, but there is a tone of regret nevertheless. As far as comparisons heard between the present and the recent past in amity, mutual assistance, and cooperation among neighbors, there is no ambivalence and no difference between male and female informants: all compare the present unfavorably.

In fact, so far as objective evidence is available, it appears that community life has declined along the dimensions informants mention. Not that the Twelve Tribes has ever been a society characterized by boundless concord, ubiquitous amity, and widespread cooperation. As we have seen, throughout the history of the Twelve Tribes, the two sections of the community, the Nine Tribes and Three Tribes confederations, have competed vigorously across a wide scope of activities, and there have been more than a few periods of sharp conflict. Still, the sections had been mainly united in a dynamic opposition that allowed the community as a whole to exist and even prosper. Sometimes the help of outside authorities, including the notably successful Mazrui rulers in the last century, was involved in overcoming the opposition when it became disruptive, but in the last century or two, a dynamic unity was usual.

As seen earlier, the confederations were made up of migrant "tribes," many of which included within themselves a variety of immigrant descent lines that, in time, became thoroughly integrated into the community through membership in one or the other of the sections, each localized in its own part of Old Town. A partial exception to this integration into the community through section membership is found for a number of families founded by men of Omani origin. These men married Swahili women and produced, over the generations, descendants who continued to emphasize their Arab identity even though in most respects they behaved in accord with Swahili, rather than Arab, culture. A majority of these "Swahili Arab," as they can be called for easy reference, families were in the Nine Tribes section, but the Three Tribes also included some. Despite their insistence on their Arab heritage, they participated in the rituals and activities of their confederations and of the community as a whole and patterned their social lives, including marriage, just as other Twelve Tribes members did.

With the ascendancy of the Busaidi in the nineteenth century, however, the internal solidarity of each of the sections was lessened by the commitment of the Swahili Arabs to the ethnically related (as they themselves understood that relationship) group and culture on Zanzibar, that is, that of the Busaidi

sultan and his retainers. I shall refer to this latter group, again for ease of reference, as "Zanzibar Arabs." It will become clear that the commitment of the Swahili Arabs of each section to the Zanzibaris has worked equally against the interests of the members of the opposite section and of members of their own section who do not share their claimed ethnic origin.

There was nothing new about the identification of the Swahili Arabs with the peoples of the Omani region. For many generations, there have been Swahili families that traced their founders' origins to the Persian Gulf area, but when the identification was used as a resource to gain political advantage beyond the community and to benefit from colonial racial policies, it had an influence both more profound and more lasting than it had ever had before. One of the earlier advantages that came to those who could prove they were of Arab descent appeared in 1910 when "Arabs" were exempted from the colonial Hut Tax, while others, including Swahili who did not go to court to prove they had Arab ancestors, were not (Salim 1973:187).

Asserting Arab Ethnicity and Its Effect on the Community

Whatever their gain by asserting it, my hypothesis is that the effect of the employment of the claim to Arab, rather than Swahili, status by some Twelve Tribes members was to disrupt the long-standing organization of the community. That is, the insistence on an Arab identity by some members of each of the sections united the remaining members of the sections with each other across section lines in opposition to those in both sections who claimed the identity. This realignment, I maintain, is a major factor in the weakening of the community in that it undermined the long-standing relationships within and between sections that had served as the basis for community structure since the Three Tribes joined the Nine Tribes on Mombasa island centuries before.

The details of the conflict between those emphasizing their Omani origins and other members of the community in both sections are tangled and baroque. In part, the complexity is based in the fact that there were, and are, Omanis who live in Old Town who are not Swahili. These families follow the Ibathi canon of Islam, the men wear beards, and their patrilateral, and in some cases even matrilateral, forebears came from Oman or elsewhere in the Persian Gulf region not more, in most cases, than two or three generations ago. These families speak Arabic in their homes, and although they associate with their Twelve Tribes neighbors, they are not considered by any community members to be Swahili.

The Swahili Arabs, however, are a different group, and none of the religious, linguistic, or descent characteristics just noted for the Omanis is true

of them. The fact that the neighboring Bantu-speaking peoples (the Giriama, Digo, and the others who are jointly referred to as "Nyika" by the Swahili) refer to both Swahili and Arabs (whether assimilated or not) as Wazomba^[2] indicates that from the outside, at least, the differences between the two are not always obvious or salient.

The distinction between the groups is not always an easy one to make from any perspective; this difficulty lies, in fact, at the heart of the conflict within the community that became serious and disruptive in the 1920s. Members of the Swahili Arab group denied the validity of any distinction between themselves and less assimilated immigrants from Oman. They, the Swahili Arabs, insisted that they were members of the same group as the Zanzibar Arabs. They claimed that this membership entitled them to the considerable privileges accorded the latter group under colonial rule.

This claim was mainly accepted by the Zanzibar Arab government and the British who advised and, in the twentieth century, succeeded them, and it won for the Swahili Arabs advantages not open to other community members. This was so, informants report, under the administration of the sultan up to 1895 and it continued under the colonial administration when the Mombasa area was administered by a special arm of the Kenya Colony government, called the "Arab Administration," whose highest officials were Zanzibar Arabs.

New Community Divisions

The Swahili Arabs took considerable pains to align themselves with the Zanzibaris in ways beyond simply asserting common ethnicity. Sh. Hyder Kindy, who was personally involved, gives a lengthy account of some of the key events concerning the assertion of the alignment in the 1920s in his Life and Politics in Mombasa (1972:26–45). The accounts I have received from three other participants in the events substantially agree with Kindy, but they, and I, do not agree that all of those he identifies as "Arabs" are, in fact, what I am calling "Zanzibar Arabs" or Omani. Some of them were Swahili Arabs, and they and their families are Twelve Tribes members with all the social, cultural, and linguistic traits appropriate to members of that community.

Under colonial rule, the political advantages of being classified as "nonnative" were substantial, and the stronger position of Arabs was clear to the Swahili as it was to everyone else. Thus, despite an initial rebuff, in 1921, many men from the Twelve Tribes, both Swahili Arabs and others, joined with resident Zanzibar Arabs in a political group called the Coast Arab Association (Salim 1973:180–187, Stroebel 1979:40) with the intention of increasing their ability to influence the colonial government (Kindy 1972:29–31). In 1927, the liwali (chief administrator) of the coast, a Zanzibar Arab

named Ali Salim (later Sir Ali) who opposed and was opposed by the association, chose it as the venue for the announcement that Twelve Tribes members were not Arabs and would not be allowed to vote for the Arab representative on the Legislative Council (ibid., 30).

This led Twelve Tribes members who did not claim Arab ethnicity to form their own political group, the Afro-Asian Association (ibid.). It also led to a high level of conflict between those who emphasized their Arab ethnicity (both Swahili and Zanzibar Arabs) and those who did not (ibid., 31–45). There was even a cessation of communitywide prayer (ibid., 31). Formerly, the men of the Nine Tribes and the men of the Three Tribes alternated the Friday noon prayer between the main Three Tribes mosque and the main Nine Tribes mosque. In 1929, however, a Three Tribes man rose and denounced "the Arabs" and advised everyone to avoid going to the Nine Tribes mosque any longer since it was where "the Arabs" prayed. This despite the Muslim injunction that the whole community pray together at least at the Friday noon prayer (ibid., 31–32).

The appeal was effective, and most of the other Twelve Tribes members, including some Nine Tribes members who did not claim Omani origins, prayed every Friday in the Three Tribes mosque. Those emphasizing their Arab connections remained in the Nine Tribes mosque, with few others joining them there (ibid.).

This breach in Nine Tribes solidarity was also a blow at the unity of the sections as such in that for the most important prayer of the week, the men of the community no longer assembled according to section divisions. There was a limited unity between members of the two sections who did not claim Arab status, but the structure that had characterized the community for centuries was seriously impaired.

"Natives" and "Nonnatives"

The strain on the community resulting from the separation of segments of the sections and the reuniting of these segments according to ethnic ties rather than section allegiance was continued and reinforced by the ability of some members to achieve what was, under British rule, the politically and economically more desirable status of Arab. The interest in doing this is seen in the fact that continued appeals from Swahili to the colonial government led, in 1934, to the ruling that

persons who could prove before a magistrate that one parent was of nonnative descent could press a claim for nonnative status, thus opening the possibility for Twelve Tribes claims. . . . Until World War II bickering continued about whether Twelve Tribes persons should be allowed to claim Arab status as nonnatives. Technically, "Swahili" were given Arab status in 1952, but relations between the two communities remained strained (Stroebel 1979:40–41).

The interest in being classified as an "Arab," that is, a nonnative, which had provided tax benefits for decades, received further impetus with the outbreak of World War II when food rationing was instituted. Those classified as natives were given coupons to buy cornmeal, while those classified as Arabs were, like Asians, allowed to buy rice (Kindy 1972:109). This was especially significant because of the meanings attached to rice and to cornmeal. For the Mombasa Swahili, eating cornmeal is inappropriate for proper group members. True Swahili of noble birth (waungwana , sing. mwungwana ), that is, those without slave forebears, simply do not eat simi , the heavy cornmeal paste^[3] eaten throughout East Africa or, at least, do not let it be known that they do. Rice is the starch suitable to waungwana, and not to have it is a degrading and shameful indication of abject poverty and/or low taste.

Informants report that even for those who received it, the rationing did not provide enough rice for it to be the dietary staple, as the Twelve Tribes members I know insist it must be. Still, being closed off from legal access to the noble grain while their fellow group members, the ones claiming Omani origin, had it was an extremely bitter experience that is remembered with rancor more than forty years later.

Crucially for the thesis being developed here, the resentment went not only to the government^[4] but also, and mainly, to the group members who claimed Omani roots. Again, this united part of the Nine Tribes with most of the Three Tribes against the Swahili Arab subgroup drawn from both.

Section Competitions

The rationing and the earlier blows to section unity affected the ability of the sections to unify and compete with one another. Nevertheless, until the early 1960s, the members of the Three Tribes and the Nine Tribes did engage in such sectionally based competitions as team card games, various sports, and marching societies (quaride ) that competed in precision of marching, elegance of uniform, and the skill of their bands. These primarily male activities were paralleled by competitions between women's dancing societies, called vyama .

Continuing until independence in 1963 was what several male informants have said was the most basic and fundamental expression of community life: the performance of a men's dance called tware and a related one called diriji . A large proportion of the men from each section, including the "Arabs," participated in these. In tware, the men from each side formed two lines, one for the Nine Tribes and one for the Three Tribes, facing one another. To the measured beat of the tambourine-like tware drum, each side attempted to outdo the other in the gleaming whiteness of their gowns and kofia (white skullcaps worn by Muslims) and in the elegance and grace of their movements in this very restrained dance.

No prize was awarded; in fact, no judgment was made. But each side assessed its own performance against that of the other and decided for itself who had been the most "noble." And "noble" is the word for this dance. It is performed in celebration of the marriage or circumcision of Twelve Tribes members but only for those who are understood to be descended from forebears all of whom were free men and women (i.e., waungwana). Moreover, only those with this sort of family background were allowed to participate in the dance.

A similar sort of dance, diriji, was also held and was seen, informants say, as another expression of community life. It, too, was restricted to those considered waungwana and was performed by the confederations in opposition to one another.

Many of the most important relations between women in the community were, like the men's dances, carried out in sectionally organized groups. This is so despite the existence of a group made up of all the "noble" old women, the wamiji (miji refers to city; wa - is the suffix for nouns referring to humans). This group acted as the ritual guardians of the community without respect to section lines, condemning improper ceremonial and ritual behavior and lending their presence to important celebrations regardless of section membership. Old men had the same title but seem not to have actively involved themselves as the women did (Stroebel 1979:80–84).

Like the wamiji, section lines were not regarded for the weddings, funerals, and circumcisions that were the center of their social life in this sexually segregated society. Women invited all community members of their gender to the rites. This was so even though those who cooperated most closely with one another in the laborious and elaborate preparations for these ceremonies were almost always from the same neighborhood and, therefore, section.

But this does not mean that sectional opposition had no part in female activities. In women's social lives, the competition between the sections came out most clearly in the women's dancing societies, or vyama (sing. chama ).^[5] There were a number of these societies, but the two main ones were based on section membership (ibid., 160–164). In a way somewhat similar to the men's marching societies, the women's competition involved elegance of costume, skill in dancing, and excellence of music between section-based groups whose members included the descendants of slaves as well as women whose forebears were understood to include only waungwana.

The women's competition went beyond those of the men. In addition to dancing skill, they also competed in the excellence and lavishness of the food presented at their dances, the elegance of their clothing and jewelry, and, especially, in the mordant wit of the songs reviling members of the competing group. These were sung at the dances and dealt with such embarrassments of the opposite section as one of its men having elephantiasis of the testicles, the pretensions to high social standing despite having a slave fore-

bear of one of its families, and the sexual indiscretion of one of its women. Men, especially older and more prestigious ones, disapproved of these women's competitions, but since they were mainly carried out within the confines of the women's separate groups whose activities were not held in the men's presence, their disapproval only kept their own wives and daughters out.

Unity Through Competition and Its End

The pervasiveness of conflict or, at least, of sharp competition, which was sometimes difficult to differentiate, between the two sections continued in a variety of forms for roughly three decades after the end of World War II. It was mainly in competition that the whole community came together. The important joint prayer on Friday was no more, but the men's dances, dirigi and, especially, tware, exhibited and affirmed some of the most prized values for men in the context of a competitive unity. The other competitions and oppositional joint appearances did not have the dignity and value-heavy significance for community coherence that tware had and the joint prayer had had, but they did bring community members together in actively functioning alignments that took in all parts of the group.

These sorts of activities, however, received a serious blow from a single, dramatic event in the early 1940s when the long-standing contests between sectionally based women's dancing groups escalated into street fighting involving the police (ibid., 177–181). This happened in some part because of changes in the women's understandings of what limits there were on their public behavior. Particularly at issue was the extent to which they were willing to be guided by the understandings men (i.e., their husbands, fathers, and brothers) had of how they should behave, especially how they could express themselves publicly.

The most prestigious men in the community had always looked on the women's societies as unacceptable expressions of tendencies in the community that they deplored: the public appearance of women, direct and open attacks on the private lives of community members, and the participation of waungwana and the descendants of slaves in common groups. They opposed the latter because the women's organizations did not practice the exclusion of those of other than "noble" birth as the men's dances did.

The riot shocked both men and women, but it did not surprise the senior men—or so some of them told me—who deplored the women's organizations and their activities from the outset. The most important consequence of the excesses of the women's dancing competitions for the future of the community was the unfavorable light it cast on all competitions, including the traditional ones between sections.

The women's riot led directly, informants have told me, to the abandon-

ment by members of the Twelve Tribes community of this whole type of competition. This is an exaggeration since, in fact, the men's dances, card games, and such went on for as long as two decades after the riot, and some of the boys' soccer teams are still sectionally based.^[6] Nevertheless, it is probably true that the riot gave all competition a more worrisome connotation. It is a matter of fact that the experience is cited forty-five years later as an example of the foolishness and danger of competition, especially between women, given the widely shared understanding that they are uncontrollably emotional.^[7]

National Politics and Its Indirect, Profound Influence

The lessened vitality of the sections was carried substantially further by one of the general understandings of the consequences of Kenya gaining independence in 1963. As many Swahili saw it, the officials of the new government were sensitive to the fact that community members had once owned slaves, and it was (and is) believed that any reminder of that should be avoided.

As has been noted, the sectionally organized men's dances at weddings and circumcisions were performed only on behalf of families and individuals whose ancestry is understood to contain no one of slave background. Similarly, no one could participate in the dances without his being understood to have no slave forebears. At least during the period of my visits to Mombasa beginning in 1975 and, according to some informants, since independence, no men's dances have been performed in Old Town.

It is noteworthy that the cessation of the dances strikes not only at the opposition of the sections and their association with highly prized male traits but also at a key foundation of the uniqueness of the community as a whole: the noble, that is, nonslave, status of all "true" community members. All Twelve Tribes members still speak of uungwana , nobility, and although the term refers to the most prized forms of behavior, manners, and character, it also applies to having an ancestry free of forebears who were wazalia (a polite term for slaves). This fundamental value is no longer publicly symbolized, and although it is still cherished and spoken of in private circles, its public demonstration in dances associated with key rituals has been abandoned.

Current State of the Community and the Section System

Although the absence of men's dances is not mentioned as a particularly important part of the phenomenon, everyone heard to comment on the general

state of the community notes a distinct decline. A number of sources for this have now been seen.

It is of considerable interest to note that two of the most important sources concern the boundaries of the community and the statuses of those who are at those boundaries. It would be more than bold to single out some particular aspect of community life or set of circumstances as sufficient to have brought about the state of the community as at least some members currently assess it. But this does not mean that particular elements cannot be singled out as having played identifiable parts. I will argue that the nature of the relations between the sections played one of these roles and that this, in turn, is shaped by the statuses of those who compose the sections.

Another important role will be shown to have been played by the nature of membership in the group. Here the issue is the group's boundary, as this is expressed in the status of group members. There are, as suggested above, two sets of individuals whose category membership presents difficulties. The first are those with known slave forebears. They are understood by group members as seeing themselves as belonging to the group but whose membership is denied by a majority of group members. The second problem set is composed of the Swahili Arabs. They are seen by a majority of community members as rightfully belonging, but in their own assignment they do not.

"Ethnic" Status and the Destruction of the Two-Section System

This last problem in status membership involves a category of people who are understood as rejecting the group member status in favor of being categorized in the Zanzibar Arab group, with the prestige and material advantages under British rule such membership entailed. The Zanzibar Arab status has implications for the social structure of the community as a whole in that it brought together the individuals who claimed it in a unity that crossed the lines bounding the two Swahili sections. Resistance to their being categorized in it united against them, again across the boundaries, members of both the sections. These status realignments affected the internal unity of the sections in such clearly manifested ways as the abandonment of the weekly prayers as part of the resistance to the Zanzibar Arab status for community members. This diminished unity lessened the ability of the sections to establish and affirm the community's vitality in the manner that had long been characteristic, namely, through competitive activity.

The division between the sections had been, in classical Gluckmanian dynamics (Gluckman, Mitchells, and Barnes 1963:1–2),^[8] a key base of social unity for the Swahili. It was in many respects destroyed by the formation of

new unities across what once had been section lines. These new "ethnic" unities produced no wide-scope solidarity because they benefited the Swahili Arabs as individuals (through better rations, jobs, and taxes), not as a group, and united their opponents from the two sections only in opposition to them. Attempts to unite the two new groupings into political action groups went on for a number of years (Kindy 1972:26–45) but disappeared after independence. They survive only as vague memories in the minds of middle-aged and elderly community members.

Another force in diminishing community vitality is the other status ambiguity mentioned above. The descendants of slaves have been closely associated with the group for generations, and not a few are widely viewed by group members as worthy embodiments of the values held by group members. However, a key identifying understanding for those with group member status is having only waungwana ("noble") forebears. To accept those whose genealogies are known to include ancestors who were not noble would be to alter the group member status radically.

An indication of how seriously this was (and is) taken is that a main reason for rejecting the proposal of marriage from a man's family, one that rated as potently as the fear that the man was a drinker of alcohol or a passive homosexual, was suspicion of his family having wazalia (nonnoble) forebears. Even if the woman's family accepted a proposal, there was the possibility that, as recently as the 1950s, the wedding ritual would be interrupted by the wamiji, the elderly women guardians of ritual and propriety who were then still active, if they believed there were genealogical irregularities.

Given the common understanding that calling attention to the fact that there were slaves in the community was at best tactless and quite possibly dangerous, holding public rituals emphasizing these differences took on a new significance. This is just the emphasis of the tware and diriji dances through their excluding all but nobles and making a central point of demonstrating the essence of "nobility" (uungwana ).

According to several community members, this is why the dances stopped. Their analysis appears to be correct. The dances were an important symbol of the community's vitality based not only in the central values concerning nobility (see chap. 7) but also in the opposition of united sections competing for commonly understood and equally prized honor. These dances, in their movements and costumes as well as through the requirements for their participants, expressed and symbolized what is surely the most commonly cited (by community members) and, probably, the most crucial values in the group, those centering around nobility. Moreover, this expression occurred with the members of each section united with his fellows in competition with the other united section in expressing their shared ideal understandings and gaining prestige in terms of them. Informants say the dances were the most "beautiful" events in public life; they seem nearly pure enactments of the com-

munity's broadest social structure, including understandings about who the occupants of its component statuses were and how they should act.

The immediate basis for the end of the performances was the exclusion of the occupants of the slave descendant status and the government resentment this was understood to incur. That I have seen few indications that this resentment is actually likely and would be an active force if it were is relatively unimportant. That it was taken so is what matters.

Nor is the end of the men's dances the only change affecting community structure. If the Swahili Arab status is more salient than the status of Nine Tribes or of Three Tribes members, the oppositional basis of the community's coherence is changed and lessened. The community continues as an interconnected structure of statuses that actually guides their members' behavior. But the scope of the expectations in the relevant statuses has been decreased, their salience has been lessened in many situations, and the identifiers have been weakened. In fact, understandings about membership in the community are much as they long were, but the symbolizing of that membership is muted and privatized and this, like the change in sectional opposition, has surely affected the nature of community life, which continues but is different.

Marriage and Community in Contemporary Mombasa

Despite the general changes and the group's diminished economic and political situation during the twentieth century, the Swahili family and sense of community continue as major forces in the lives of group members. My informants and the literature (e.g., Stroebel 1979:80–94, Prins 1967:76–83) agree in indicating that for at least several decades, kinship, like community, is neither as broad in scope nor as powerful in directing activity as it was in the recent past. Still, ties based on both affect much of what most group members do most of the time. The Swahili are an urban people, and they live in an unquestionably rapidly changing environment, but they are not deracinated, their nuclear families are not isolated, and their senses of identity are still, for almost all group members, firmly rooted in being Swahili.

In later chapters, marriage and the nuclear family receive a good deal of attention, but from the perspective of community structure, it is useful to bear in mind that almost all marriages within this community are with fellow community members. Men sometimes marry Mombasa women from other ethnic groups, but when they do this—and this is true now as it has been for at least the greater part of this century—the marriage is kept secret, especially from the man's "main" wife.^[9] This wife is almost invariably not only a Mombasa Swahili but, nearly as frequently, also a member of the same section as her husband and, in more than a quarter of the cases for which I have data, either a patrilateral or a matrilateral cousin.

In recent years, according to some informants, ties beyond the nuclear family have weakened, but if this is so, it is so only relative to what must have been quite powerful bonds. Households often contain kin in addition to nuclear family members, and if the sharing of houses by married brothers and their wives and families is rare, as it is, this is not a new development; a separate house for each nuclear family, especially after a number of children have been born, is generally thought desirable and was so viewed as far back as informants can remember.

It is difficult to assess whether the Swahili are virilocal or uxorilocal since crowding and the lack of availability of house sites makes it difficult for couples to live where they and their kin wish. Some informants say it is better for a newly married couple to live near the groom's family, and others (fewer) say it is better to live near the wife's family. Since an accurate census is quite impossible to carry out,^[10] I can only say that if I have an impression as to where married couples live, it is that there is some tendency to locate near the wife's family. My census data (see table 1, chap. 4), as spotty and thin as they unavoidably are, indicate that couples always live in their section's neighborhood if they are from the same one (as they almost always are) and if, as the overwhelming majority still do, they live in Old Town rather than in non-Swahili parts of the city or in a distant town or nation.

Women from the group occasionally now, and in the past as well, marry men from other Swahili communities along the coast, and a few marry Arabs, either from the Persian Gulf or from the community of unassimilated Arabs in Mombasa. Even less often, they marry whites who have converted to Islam; I know of three cases, two to Europeans and one to an American. These two rare sorts of exceptions aside, however, the group's women never marry outside their own community for their first marriage.

Parents are concerned that their children may marry outsiders and, even more frightening, Christians, but this has not happened in any of the forty-six marriages since 1975 which I have data on. If there are marriages outside the community, it is in the pattern of men taking "secret" wives (all of them, so far as I can tell, Muslims) that has been part of the marriage practices followed for as many generations as informants' accounts go back.

"Clans" and Other Designations Wrongly or Rarely Used

Turning to the family and kinship, it was noted above that the community's constituent sections, or confederations, are composed of what is glossed as "tribes" (mataifa ) and that these, in turn, are made up of patrilateral descent groups (mbari ), called "clans" in some historians' accounts (Berg 1968:40–42, Pouwells 1987:79, 84 passim). The use of the term "clan" for any part

of this society, however, is likelier to produce misunderstanding than to illuminate its social life.

The kin groupings that compose the tribes were not based on strict patrilineal descent, did not in all cases believe in a common ancestor, were not exogamous, and were only localized in the sense that all the members of a confederation or section lived in the same area without, however, any residential distinction by descent groups. The only basis for using the term "clan," in fact, is that English-speaking group members use it.

Looking at this usage, however, strongly suggests that the term is applied loosely to "kin" or "fairly close kin" rather than to any sort of unilineal descent group. For example, an informant said "the clan" was coming to a gathering at his house, and the people who came included not only his patrilateral kin but also some related through his mother and some through his wife (who was not his cousin). When I asked him if he considered them all to be "his clan," he said that he did. This is quite consistent for the use of "clan" as described by Prins (1967:80–81), whose reservations are in accord with those noted here.

The two words for groupings of kin that were used in the fairly recent past but seem not to be currently in use are nyumba and mlango. "Nyumba" is most readily glossed as "house" and is used for the nuclear family members and their siblings and parents. "Mlango" means door in Swahili and, in this usage, refers to the broadly conceived category of kin and, sometimes, affines related through either parent. It was never used to refer to a social group of any kind. It was, rather, a category used to place an individual or nuclear family in a context of kin and affines. Prins notes that the term "mlango" is an alternative to "clan" or mbari. Kindy gives a historically useful list of the names of all the mbari making up each of the tribes (1972:49–51; see also Berg 1968:40–41).

There are hints in my informants' discussions that mbari were once the basis for joint activities including warfare and dances, but if this did occur in the past, it has not done so for most of this century. Many informants, including some of those who know their "tribe," do not know which mbari they belong to. Even in the earlier period when mbari was widely known, it appears that it was not the primary framework for close relationships beyond the household. This framework, especially for women and children, was and still is provided by the mtaa (pl. mitaa ), or neighborhood.

Residence

Residence was more nearly uxorilocal than virilocal, and although patrilateral cousin (both cross and parallel) marriage was fairly common, so was matrilateral cousin marriage. Marriage to unrelated individuals, mainly of the

same tribe and almost always of the same confederation, accounted for more than half of all marriages. In fact, informants say, marriage with people from the same mtaa was more common than any other sort, and such marriages could be with cousins on either side or with neighbors who had no traceable kin relationship. As a thoughtful and intelligent informant put it, "The mtaa was like a group of relatives. The neighbors could be related on either side or maybe they weren't obviously related, but they had been marrying each other for generations, so they were all related."

The mtaa is still socially important as the main basis for the young people's groups of friends (still completely separated according to sex), the visiting and mutual assistance groups of women, and the men's barazas .^[11] There are signs of the weakening of the neighborhoods of some Swahili families living outside Old Town and an ever-increasing number of houses in the old neighborhoods being taken over by members of other ethnic groups (see Swartz 1983:36–37), but as of the mid-1980s, they remain the most vital element in community life beyond that based in close ties of kinship.

Like the mbari, the tribe has not been an active group since the 1960s. Few people, and none I have met who are under 35 or so, know to what "tribe" they belong. The confederations of tribes remain symbolically important, however, and although they carry out no social activity, I have not talked to a single community member who is unaware of what confederation he or she belongs to. There is no general Swahili term that can be translated as "confederation" or "section," but it is the most frequently heard wide-scope identifier for community members, so that one hears that someone is "a Three Tribes member" or "a Nine Tribes member" on the rare occasions when broad social placement is at issue.

Outside Contacts and Community Importance

Despite the continuation of virtual community endogamy, there is more rather close contact with non-Swahili Mombasans than there was in the past. The secularization of the community increased under the British as the Swahili became a smaller and smaller proportion of the city's population and were increasingly integrated into an economic system where most of the employed men (very few women worked for wages outside the home until the last few years when a small number got paying jobs) spent their days working with mainly Christian members of other ethnic groups.

This has not led to any movement away from Islam. Almost any conversation, even a brief one, leads to at least one reference to the Koran, or to God, or to the holy laws. Religious courts with Swahili judges (kadhi ) learned in Koranic law are invariably used by community members for difficult

domestic problems and civil disputes among themselves. Civil courts are involved only when there are serious crimes. The boundaries of the community, in fact, can be traced by examining the court—religious or secular—that disputants use (Swartz 1978).

The Swahili have always insisted on Islamic education for their children of both sexes, and I know of no one in the community who did not learn to read the Koran and to pray in the chuo (religious school for children). Secular schools in the area were initially, in the nineteenth century, mission operated, Christian oriented, and designed to produce converts, so it is little wonder that the Swahili (and other Muslims) avoided them. Eventually, however, by the 1930s, the economic importance of a secular education together with concessions to Muslim interests (including teaching the Koran and Arabic to students in the schools the Swahili attended) led to a substantial proportion of the community sending their sons and, later, even some of their daughters to government-run schools including secondary schools (Salim 1973:146–156) despite the high cost of tuition.

Diminished Prosperity

In the period following World War I, the Mombasa Swahili began to be acutely concerned about their economic situation and prospects and about the future of their group and their religion. As Salim (ibid., 179) says,

The aim to raise their status . . . was widespread. . . . It was the product of the economic plight which the new generation found itself heir to after their fathers' rickety system of trade and cultivation . . . had finally crumbled under the impact of the abolition of slavery and the advent of the railway, and after a good deal of land had been lost to government . . . or sold to foreign speculators [i.e., Indians and Europeans] in order to maintain a semblance of the old standard of living.

The Swahili community does continue to "maintain a semblance of the old standard of living." It appears that almost half the households have a servant, and color television sets and videocassette recorders are common (even though each one costs half a year's salary for a junior civil servant). Community members are well dressed, and informants say that as many as one extended family in three has a member who has made the expensive hajj to Mecca.

Compared to the Indians and Hadhrami Arabs who live among them in Old Town, however, the Swahili are not well off. Only one major business in Mombasa is owned by a community member, and most men are employed as civil servants, teachers, small shop owners, clerks, managers, and middle executives. A number of younger men have been employed in Saudi Arabia

and the Persian Gulf area, and their remittances were important sources of income for a substantial number of families through the mid-1980s when this diminished. Mombasa Swahili have never done manual work save as fishermen, and this continues to be true, but wealth on a par with their non-Swahili neighbors has eluded them since the early part of the century and continues to do so.

The Swahili are well-to-do in comparison to many other ethnic groups in Mombasa, but compared to their past—and their memory of their past—as well as to their more affluent Indian and Arab neighbors, their situation is a difficult one. Nor is this only psychological. There are genuinely poor families in the community, and serious concerns about money and debts are more often characteristic of families than are feelings of financial security.

Less Expensive Life-Crises Rituals

Partly as a result of their limited economic resources, some of the opulent life-crises rituals that once characterized the community (Prins 1967:104–105, Stroebel 1979:8–13) have paled and others have vanished. Group members still talk about these rituals with a mixture of nostalgia (for the glory) and sorrow (for the great expense).

Funerals, especially those for prominent persons, are still attended by hundreds upon hundreds, but the spendthrift days of feeding this multitude for days and weeks have come to an end. Birth is now the occasion for only a relatively modest celebration by a few score women relatives and neighbors. Circumcision, once an occasion for a major feast, is no longer publicly celebrated. Weddings were lavish occasions to which the entire community was invited as recently as the late 1970s, but now they are less opulent and much smaller, and attendance by more than one hundred is rather unusual.^[12]

As of the 1980s, some weddings and other life crises are marked only by a maulidi (reading of the life of the Prophet followed by light refreshments) to which only kin and neighbors are invited. The expenses involved in the grand rituals were enormous: as much as five years' earnings were spent on a wedding or a funeral. As one walks through Old Town, one is shown houses that families mortgaged and lost to pay for these rituals. But expense is not the only reason these rituals have declined in frequency and opulence.

One is that Mombasa, including the Old Town section, has gone the way of cities everywhere: crime has increased, while generally acceptable behavior has decreased. Robberies and murders, although still rare by the standards of Nairobi and most American cities, have increased sharply since the latter years of the 1970s as rural migrants streamed into the city. Now many community members have well-founded uneasiness about walking through the streets of Old Town at night. Because of this, attendance at prayer in the

mosques at the predawn, alfaquiri, prayer and the nighttime, isha, prayers has declined, with more men (women do not enter Swahili mosques) praying at home. The celebration of rituals was and is held in the streets adjacent to the house of the family involved, and as crime increases, community members are less and less enthusiastic about being out of their homes after dark.

Further difficulty comes from that fact that wahuni, rowdy boys and young men, some of them Swahili, are far more plentiful than they were and, many say, more unrestrained in their behavior. Uneasiness about celebrations being ruined by fighting and vandalism has a substantial basis in experience. It was traditional for the groom to be escorted to his bride's house for the wedding night and for his entrance to be opposed by neighborhood boys and young men. After disarranging his clothes and making some noise, they let him in—sometimes after being given a token gift of money. Lately, however, these boys and young men have sometimes become rowdy and even dangerous, seriously beating the groom and his escorts and breaking windows and furniture in the house of the bride's family.

Even more important, perhaps, is the decline in the community discussed earlier. Men are heard to say that money spent on lavish entertainment is wasted and that people are only impressed with the foolishness of someone who provides an elaborate wedding or other ritual for the community at large. Men have long taken this view but have yielded to their wives' wishes to stage impressive rituals (Swartz 1982b, 1983).

It may be that women are no longer as motivated as they were just a decade ago to spend an important part of their family's wealth on one or two rituals as well as slightly less able to influence their husbands. The reasons for this are complex and will be addressed in chapter 10. For the moment, it is enough to say that the social relations that played the central role in the women wanting to hold expensive rituals and, at the same time, provided them a source of freedom in dealing with their husbands have decreased in importance for them as the community's overall integration has declined.

This is not to say that the rituals have stopped. They continue, if on a reduced scale, and give every evidence of accomplishing the ends of life-crises rites as set out long ago by Van Gennep. Like the community, they are still effective but reduced in social scale and toned down in their opulence.

The culture of the Mombasa Swahili is still effective and still guides the activity of most community members, most of the time. The center of their lives has always focused around kinship and the household group, and whatever changes have taken place and are occurring have not and do not alter that fact.